[#1] Guide LLM to utilize MySQL FULLTEXT search indexes in app.py prompt

app.py

@@ -801,8 +801,13 @@ DATABASE SCHEMA (MySQL database 'food_db'):
 TOOL TAGS (Generate these inside your response when you need details):
 1. To run a read-only SELECT query, write:
    <sql_query>SELECT ...</sql_query>
-   Make sure to join tables on `code`. Optimize queries using LIKE. For example:
-   <sql_query>SELECT c.product_name, m.proteins_100g FROM products_core c JOIN products_macros m ON c.code = m.code WHERE c.product_name LIKE '%apple%' LIMIT 5;</sql_query>
+   Make sure to join tables on `code`. Optimize queries using the FULLTEXT indexes on `products_core` for maximum speed.
+   Available FULLTEXT indexes:
+   * `idx_prod_name` on `product_name`
+   * `idx_ing_text` on `ingredients_text`
+   * `idx_prod_ing` on `product_name, ingredients_text`
+   Use MATCH() AGAINST('+word' IN BOOLEAN MODE) for searching. For example:
+   <sql_query>SELECT c.product_name, m.proteins_100g FROM products_core c JOIN products_macros m ON c.code = m.code WHERE MATCH(c.product_name) AGAINST('+apple' IN BOOLEAN MODE) LIMIT 5;</sql_query>
    
 2. To run a web search for information not in the database, write:
    <web_search>query</web_search>